Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion benchmark/buffers/buffer-bytelength-string.js
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ const common = require('../common');
const bench = common.createBenchmark(main, {
type: ['one_byte', 'two_bytes', 'three_bytes',
'four_bytes', 'latin1'],
encoding: ['utf8', 'base64'],
encoding: ['utf8', 'base64', 'latin1', 'hex'],
repeat: [1, 2, 16, 256], // x16
n: [4e6],
});
Expand Down
31 changes: 31 additions & 0 deletions benchmark/buffers/buffer-copy-bytes-from.js
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
'use strict';

const common = require('../common.js');

const bench = common.createBenchmark(main, {
type: ['Uint8Array', 'Uint16Array', 'Uint32Array', 'Float64Array'],
len: [64, 256, 2048],
partial: ['none', 'offset', 'offset-length'],
n: [6e5],
});

function main({ n, len, type, partial }) {
const TypedArrayCtor = globalThis[type];
const src = new TypedArrayCtor(len);
for (let i = 0; i < len; i++) src[i] = i;

let offset;
let length;
if (partial === 'offset') {
offset = len >>> 2;
} else if (partial === 'offset-length') {
offset = len >>> 2;
length = len >>> 1;
}

bench.start();
for (let i = 0; i < n; i++) {
Buffer.copyBytesFrom(src, offset, length);
}
bench.end(n);
}
1 change: 1 addition & 0 deletions benchmark/buffers/buffer-fill.js
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@ const bench = common.createBenchmark(main, {
'fill("t")',
'fill("test")',
'fill("t", "utf8")',
'fill("t", "ascii")',
'fill("t", 0, "utf8")',
'fill("t", 0)',
'fill(Buffer.alloc(1), 0)',
Expand Down
2 changes: 1 addition & 1 deletion benchmark/buffers/buffer-indexof.js
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ const searchStrings = [

const bench = common.createBenchmark(main, {
search: searchStrings,
encoding: ['undefined', 'utf8', 'ucs2'],
encoding: ['undefined', 'utf8', 'ascii', 'latin1', 'ucs2'],
type: ['buffer', 'string'],
n: [5e4],
}, {
Expand Down
2 changes: 1 addition & 1 deletion benchmark/buffers/buffer-tostring.js
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
const common = require('../common.js');

const bench = common.createBenchmark(main, {
encoding: ['', 'utf8', 'ascii', 'latin1', 'hex', 'UCS-2'],
encoding: ['', 'utf8', 'ascii', 'latin1', 'hex', 'base64', 'base64url', 'UCS-2'],
args: [0, 1, 3],
len: [1, 64, 1024],
n: [1e6],
Expand Down
87 changes: 55 additions & 32 deletions lib/buffer.js
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,6 @@ const {
TypedArrayPrototypeGetByteOffset,
TypedArrayPrototypeGetLength,
TypedArrayPrototypeSet,
TypedArrayPrototypeSlice,
TypedArrayPrototypeSubarray,
Uint8Array,
} = primordials;
Expand Down Expand Up @@ -383,28 +382,42 @@ Buffer.copyBytesFrom = function copyBytesFrom(view, offset, length) {
return new FastBuffer();
}

const byteLength = TypedArrayPrototypeGetByteLength(view);

if (offset !== undefined || length !== undefined) {
if (offset !== undefined) {
validateInteger(offset, 'offset', 0);
if (offset >= viewLength) return new FastBuffer();
} else {
offset = 0;
}

let end;
if (length !== undefined) {
validateInteger(length, 'length', 0);
end = offset + length;
// The old code used TypedArrayPrototypeSlice which clamps internally.
end = MathMin(offset + length, viewLength);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you add a comment to why we have this change now?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, the clamping is required because we switched from TypedArrayPrototypeSlice (which clamped silently)

} else {
end = viewLength;
}

view = TypedArrayPrototypeSlice(view, offset, end);
if (end <= offset) return new FastBuffer();

const elementSize = byteLength / viewLength;
const srcByteOffset = TypedArrayPrototypeGetByteOffset(view) +
offset * elementSize;
const srcByteLength = (end - offset) * elementSize;

return fromArrayLike(new Uint8Array(
TypedArrayPrototypeGetBuffer(view),
srcByteOffset,
srcByteLength));
}

return fromArrayLike(new Uint8Array(
TypedArrayPrototypeGetBuffer(view),
TypedArrayPrototypeGetByteOffset(view),
TypedArrayPrototypeGetByteLength(view)));
byteLength));
};

// Identical to the built-in %TypedArray%.of(), but avoids using the deprecated
Expand Down Expand Up @@ -551,14 +564,15 @@ function fromArrayBuffer(obj, byteOffset, length) {
}

function fromArrayLike(obj) {
if (obj.length <= 0)
const len = obj.length;
if (len <= 0)
return new FastBuffer();
if (obj.length < (Buffer.poolSize >>> 1)) {
if (obj.length > (poolSize - poolOffset))
if (len < (Buffer.poolSize >>> 1)) {
if (len > (poolSize - poolOffset))
createPool();
const b = new FastBuffer(allocPool, poolOffset, obj.length);
const b = new FastBuffer(allocPool, poolOffset, len);
TypedArrayPrototypeSet(b, obj, 0);
poolOffset += obj.length;
poolOffset += len;
alignPool();
return b;
}
Expand Down Expand Up @@ -732,11 +746,7 @@ const encodingOps = {
write: asciiWrite,
slice: asciiSlice,
indexOf: (buf, val, byteOffset, dir) =>
indexOfBuffer(buf,
fromStringFast(val, encodingOps.ascii),
byteOffset,
encodingsMap.ascii,
dir),
indexOfString(buf, val, byteOffset, encodingsMap.ascii, dir),
},
base64: {
encoding: 'base64',
Expand Down Expand Up @@ -1118,7 +1128,9 @@ function _fill(buf, value, offset, end, encoding) {
value = 0;
} else if (value.length === 1) {
// Fast path: If `value` fits into a single byte, use that numeric value.
if (normalizedEncoding === 'utf8') {
// ASCII shares this branch with utf8 since code < 128 covers the full
// ASCII range; anything outside falls through to C++ bindingFill.
if (normalizedEncoding === 'utf8' || normalizedEncoding === 'ascii') {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Currently, ascii behaves exactly like latin1
Unsure if by design or accidentally

Copy link
Contributor Author

@thisalihassan thisalihassan Feb 18, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes, this is safe by design I am just extending the existing single byte numeric optimization to cover ASCII, since the guard already constrains it to the valid ASCII range.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please update comment to why we can add ASCII here.

const code = StringPrototypeCharCodeAt(value, 0);
if (code < 128) {
value = code;
Expand Down Expand Up @@ -1168,29 +1180,30 @@ function _fill(buf, value, offset, end, encoding) {
}

Buffer.prototype.write = function write(string, offset, length, encoding) {
const len = this.length;
// Buffer#write(string);
if (offset === undefined) {
return utf8Write(this, string, 0, this.length);
return utf8Write(this, string, 0, len);
}
// Buffer#write(string, encoding)
if (length === undefined && typeof offset === 'string') {
encoding = offset;
length = this.length;
length = len;
offset = 0;

// Buffer#write(string, offset[, length][, encoding])
} else {
validateOffset(offset, 'offset', 0, this.length);
validateOffset(offset, 'offset', 0, len);

const remaining = this.length - offset;
const remaining = len - offset;

if (length === undefined) {
length = remaining;
} else if (typeof length === 'string') {
encoding = length;
length = remaining;
} else {
validateOffset(length, 'length', 0, this.length);
validateOffset(length, 'length', 0, len);
if (length > remaining)
length = remaining;
}
Expand All @@ -1208,9 +1221,10 @@ Buffer.prototype.write = function write(string, offset, length, encoding) {
};

Buffer.prototype.toJSON = function toJSON() {
if (this.length > 0) {
const data = new Array(this.length);
for (let i = 0; i < this.length; ++i)
const len = this.length;
if (len > 0) {
const data = new Array(len);
for (let i = 0; i < len; ++i)
data[i] = this[i];
return { type: 'Buffer', data };
}
Expand Down Expand Up @@ -1253,45 +1267,53 @@ function swap(b, n, m) {
}

Buffer.prototype.swap16 = function swap16() {
// For Buffer.length < 128, it's generally faster to
// For Buffer.length <= 32, it's generally faster to
// do the swap in javascript. For larger buffers,
// dropping down to the native code is faster.
if (!isUint8Array(this))
throw new ERR_INVALID_ARG_TYPE('this', ['Buffer', 'Uint8Array'], this);
const len = this.length;
if (len % 2 !== 0)
throw new ERR_INVALID_BUFFER_SIZE('16-bits');
if (len < 128) {
if (len <= 32) {
for (let i = 0; i < len; i += 2)
swap(this, i, i + 1);
return this;
}
return _swap16(this);
_swap16(this);
Copy link
Member

@ChALkeR ChALkeR Mar 5, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what would happen if it's called on not a TypeArrayView now?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i.e. Buffer.prototype.swap16.apply(new Array(128))

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch, I will add a JS side guard

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

would that be good approach @ChALkeR ?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It changes the method to not being callable on a typedarray that is not an uint8array, but the previous behavior was inconsistent in that case anyway as the js part swapped elements and the src part swapped bytes

So I think explicitly blocking non-uint8arr should be fine and not a semver-major?
(perhaps cc @nodejs/lts just in case)

return this;
};

Buffer.prototype.swap32 = function swap32() {
// For Buffer.length < 192, it's generally faster to
// For Buffer.length <= 32, it's generally faster to
// do the swap in javascript. For larger buffers,
// dropping down to the native code is faster.
if (!isUint8Array(this))
throw new ERR_INVALID_ARG_TYPE('this', ['Buffer', 'Uint8Array'], this);
const len = this.length;
if (len % 4 !== 0)
throw new ERR_INVALID_BUFFER_SIZE('32-bits');
if (len < 192) {
if (len <= 32) {
for (let i = 0; i < len; i += 4) {
swap(this, i, i + 3);
swap(this, i + 1, i + 2);
}
return this;
}
return _swap32(this);
_swap32(this);
return this;
};

Buffer.prototype.swap64 = function swap64() {
// For Buffer.length < 192, it's generally faster to
// For Buffer.length < 48, it's generally faster to
// do the swap in javascript. For larger buffers,
// dropping down to the native code is faster.
if (!isUint8Array(this))
throw new ERR_INVALID_ARG_TYPE('this', ['Buffer', 'Uint8Array'], this);
const len = this.length;
if (len % 8 !== 0)
throw new ERR_INVALID_BUFFER_SIZE('64-bits');
if (len < 192) {
if (len < 48) {
for (let i = 0; i < len; i += 8) {
swap(this, i, i + 7);
swap(this, i + 1, i + 6);
Expand All @@ -1300,7 +1322,8 @@ Buffer.prototype.swap64 = function swap64() {
}
return this;
}
return _swap64(this);
_swap64(this);
return this;
};

Buffer.prototype.toLocaleString = Buffer.prototype.toString;
Expand Down
57 changes: 44 additions & 13 deletions src/node_buffer.cc
Original file line number Diff line number Diff line change
Expand Up @@ -1050,7 +1050,7 @@ void IndexOfString(const FunctionCallbackInfo<Value>& args) {
needle_length,
offset,
is_forward);
} else if (enc == LATIN1) {
} else if (enc == ASCII || enc == LATIN1) {
uint8_t* needle_data = node::UncheckedMalloc<uint8_t>(needle_length);
if (needle_data == nullptr) {
return args.GetReturnValue().Set(-1);
Expand Down Expand Up @@ -1200,31 +1200,59 @@ int32_t FastIndexOfNumber(Local<Value>,
static CFunction fast_index_of_number(CFunction::Make(FastIndexOfNumber));

void Swap16(const FunctionCallbackInfo<Value>& args) {
Environment* env = Environment::GetCurrent(args);
THROW_AND_RETURN_UNLESS_BUFFER(env, args[0]);
DCHECK(args[0]->IsArrayBufferView());
SPREAD_BUFFER_ARG(args[0], ts_obj);
CHECK(nbytes::SwapBytes16(ts_obj_data, ts_obj_length));
args.GetReturnValue().Set(args[0]);
}

void FastSwap16(Local<Value> receiver,
Local<Value> buffer_obj,
// NOLINTNEXTLINE(runtime/references)
FastApiCallbackOptions& options) {
TRACK_V8_FAST_API_CALL("buffer.swap16");
HandleScope scope(options.isolate);
SPREAD_BUFFER_ARG(buffer_obj, ts_obj);
CHECK(nbytes::SwapBytes16(ts_obj_data, ts_obj_length));
}
Comment on lines +1208 to +1216
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These fast callbacks are non-identical to the conventional callbacks they shadow.

  • The existing callbacks validate their argument and throws to JS if invalid, whereas your fast callbacks hard-crash the process. It might be better to validate in the JS layer, then use the same unwrapping logic on both sides.
  • Your fast callback cannot have a different return convention to the conventional callback. You will need to remove the return value from the conventional callback.

Copy link
Member

@ChALkeR ChALkeR Mar 5, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Was removing the validation instead of moving it to js intentional?

Also fast paths can keep the validation and fall back to slow i think, so we can still validate on the src side?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also fast paths can keep the validation and fall back to slow i think, so we can still validate on the src side?

This is outdated, the fast API doesn't use fallback any more (since Node.js v23.x).

However, any validation in a JS wrapper should be shadowed by a CHECK or DCHECK in the C++ binding.

Comment on lines +1208 to +1216
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fast callbacks should include debug tracking and call tests.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for sharing these I will update the code


static CFunction fast_swap16(CFunction::Make(FastSwap16));

void Swap32(const FunctionCallbackInfo<Value>& args) {
Environment* env = Environment::GetCurrent(args);
THROW_AND_RETURN_UNLESS_BUFFER(env, args[0]);
DCHECK(args[0]->IsArrayBufferView());
SPREAD_BUFFER_ARG(args[0], ts_obj);
CHECK(nbytes::SwapBytes32(ts_obj_data, ts_obj_length));
args.GetReturnValue().Set(args[0]);
}

void FastSwap32(Local<Value> receiver,
Local<Value> buffer_obj,
// NOLINTNEXTLINE(runtime/references)
FastApiCallbackOptions& options) {
TRACK_V8_FAST_API_CALL("buffer.swap32");
HandleScope scope(options.isolate);
SPREAD_BUFFER_ARG(buffer_obj, ts_obj);
CHECK(nbytes::SwapBytes32(ts_obj_data, ts_obj_length));
}

static CFunction fast_swap32(CFunction::Make(FastSwap32));

void Swap64(const FunctionCallbackInfo<Value>& args) {
Environment* env = Environment::GetCurrent(args);
THROW_AND_RETURN_UNLESS_BUFFER(env, args[0]);
DCHECK(args[0]->IsArrayBufferView());
SPREAD_BUFFER_ARG(args[0], ts_obj);
CHECK(nbytes::SwapBytes64(ts_obj_data, ts_obj_length));
args.GetReturnValue().Set(args[0]);
}

void FastSwap64(Local<Value> receiver,
Local<Value> buffer_obj,
// NOLINTNEXTLINE(runtime/references)
FastApiCallbackOptions& options) {
TRACK_V8_FAST_API_CALL("buffer.swap64");
HandleScope scope(options.isolate);
SPREAD_BUFFER_ARG(buffer_obj, ts_obj);
CHECK(nbytes::SwapBytes64(ts_obj_data, ts_obj_length));
}

static CFunction fast_swap64(CFunction::Make(FastSwap64));

static void IsUtf8(const FunctionCallbackInfo<Value>& args) {
Environment* env = Environment::GetCurrent(args);
CHECK_EQ(args.Length(), 1);
Expand Down Expand Up @@ -1622,9 +1650,9 @@ void Initialize(Local<Object> target,
SetMethodNoSideEffect(
context, target, "createUnsafeArrayBuffer", CreateUnsafeArrayBuffer);

SetMethod(context, target, "swap16", Swap16);
SetMethod(context, target, "swap32", Swap32);
SetMethod(context, target, "swap64", Swap64);
SetFastMethod(context, target, "swap16", Swap16, &fast_swap16);
SetFastMethod(context, target, "swap32", Swap32, &fast_swap32);
SetFastMethod(context, target, "swap64", Swap64, &fast_swap64);

SetMethodNoSideEffect(context, target, "isUtf8", IsUtf8);
SetMethodNoSideEffect(context, target, "isAscii", IsAscii);
Expand Down Expand Up @@ -1693,8 +1721,11 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
registry->Register(IndexOfString);

registry->Register(Swap16);
registry->Register(fast_swap16);
registry->Register(Swap32);
registry->Register(fast_swap32);
registry->Register(Swap64);
registry->Register(fast_swap64);

registry->Register(IsUtf8);
registry->Register(IsAscii);
Expand Down
Loading
Loading